Sebastes schlegelii – environmental adaptation of thehatching enzyme and evolutionary aspects of formation of the pseudogene Mari Kawaguchi1, Masahiro Nakagawa2, Tsutomu Noda3, Norio Yos
Trang 1Sebastes schlegelii – environmental adaptation of the
hatching enzyme and evolutionary aspects of formation of the pseudogene
Mari Kawaguchi1, Masahiro Nakagawa2, Tsutomu Noda3, Norio Yoshizaki4, Junya Hiroi5, Mutsumi Nishida6, Ichiro Iuchi1and Shigeki Yasumasu1
1 Life Science Institute, Sophia University, Tokyo, Japan
2 National Center for Stock Enhancement, Fisheries Research Agency, Goto Station, Nagasaki, Japan
3 National Center for Stock Enhancement, Fisheries Research Agency, Miyako Station, Iwate, Japan
4 Department of Animal Resource Production, United Graduate School of Agricultural Science, Gifu University, Japan
5 Department of Anatomy, St Marianna University School of Medicine, Kawasaki, Japan
6 Ocean Research Institute, University of Tokyo, Japan
At the time of hatching of oviparous fish embryos, the
hatching enzyme is secreted from hatching gland cells
of the embryos to digest the egg envelope (chorion) [1–
3] The hatching enzyme cDNAs have been cloned
from embryos of various oviparous fish species, such
as medaka (Oryzias latipes) [4], zebrafish (Danio rerio) [5], masu salmon (Oncorhynchus masou) [5], yellow-tailed damsel (Chrysiptera parasema) [6], Japanese eel
Keywords
aberrant splicing; adaptation; astacin family
metalloprotease; hatching enzyme;
pseudogene
Correspondence
S Yasumasu, Life Science Institute, Sophia
University, 7-1 Kioi-cho, Chiyoda-ku, Tokyo
102-8554, Japan
Fax: +81 3 3238 3393
Tel: +81 3 3238 4263
E-mail: s-yasuma@hoffman.cc.sophia.ac.jp
Database
The nucleotide sequence data have been
submitted to the DDBJ ⁄ EMBL ⁄ GenBank
nucleotide sequence databases under the
accession numbers AB353099–AB353111
(Received 17 February 2008, revised 25
March 2008, accepted 1 April 2008)
doi:10.1111/j.1742-4658.2008.06427.x
The hatching enzyme of oviparous euteleostean fishes consists of two metalloproteases: high choriolytic enzyme (HCE) and low choriolytic enzyme (LCE) They cooperatively digest the egg envelope (chorion) at the time of embryo hatching In the present study, we investigated the hatching
of embryos of the ovoviviparous black rockfish Sebastes schlegelii The chorion-swelling activity, HCE-like activity, was found in the ovarian fluid carrying the embryos immediately before the hatching stage Two kinds of HCE were partially purified from the fluid, and the relative molecular masses of them matched well with those deduced from two HCE cDNAs, respectively, by MALDI-TOF MS analysis On the other hand, LCE cDNAs were cloned; however, the ORF was not complete These results suggest that the hatching enzyme is also present in ovoviviparous fish, but
is composed of only HCE, which is different from the situation in other oviparous euteleostean fishes The expression of the HCE gene was quite weak when compared with that of the other teleostean fishes Considering that the black rockfish chorion is thin and fragile, such a small amount of enzyme would be enough to digest the chorion The black rockfish hatch-ing enzyme is considered to be well adapted to the natural hatchhatch-ing envi-ronment of black rockfish embryos In addition, five aberrant spliced LCE cDNAs were cloned Several nucleotide substitutions were found in the splice site consensus sequences of the LCE gene, suggesting that the prod-ucts alternatively spliced from the LCE gene are generated by the muta-tions in intronic regions responsible for splicing
Abbreviations
DIG, digoxigenin; Ga, Gasterosteus aculeatus; HCE, high choriolytic enzyme; Hh, Helicolenus hilgendorfi; LCE, low choriolytic enzyme; MCA, 7-amino-4-methylcoumarin; MYA, million years ago; Sg, Setarches guentheri; Ss, Sebastes schlegelii.
Trang 2(Anguilla japonica) [7], Fundulus heteroclitus [8], ayu
(Plecoglossus altivelis altivelis) [9] and fugu
(Taki-fugu rubripes) [10] Among them, the medaka enzymes
have been studied comprehensively The hatching
enzyme is composed of two proteases: high choriolytic
enzyme (HCE, choriolysin H, EC 3.4.24.67) and low
choriolytic enzyme (LCE, choriolysin L, EC 3.4.24.66)
They cooperatively digest the chorion; HCE swells the
chorion by its limited proteolytic action, and then
LCE digests the swollen chorion completely [11–13]
They act at the same time, and efficient, complete
digestion was observed at natural hatching Both
enzymes belong to the astacin family of
metallo-proteases [14]
Unlike oviparous fish embryos, ovoviviparous fish
embryos grow and hatch within the maternal body
and are then delivered from the body At the time of
ovoviviparous fish hatching, it has been unclear
whether the hatching enzyme is secreted from hatching
gland cells to digest the chorion In this study, we
observed the embryo hatching of the ovoviviparous
black rockfish Sebastes schlegelii, which is a member
of the Scorpaeniformes within the Euteleostei [15] The
hatching enzyme was identified from ovarian fluids of
the black rockfish, and the cDNAs and the genes for
the hatching enzyme were cloned from the embryos
Results
Detection of metalloprotease activity in ovarian
fluid
We expected that enzymes secreted from ovoviviparous
fish embryos (hatching enzymes) would be present in
the ovarian fluid after the embryos hatched Ovarian
fluid was collected from the ovarian cavity, and its
proteolytic activity was examined using several
sub-strates added in isotonic saline (0.128 m NaCl, similar
to the natural hatching environment of embryos in the
ovarian cavity) The teleostean hatching enzymes are
generally known to belong to the astacin family of
metalloproteases, and they are inactivated by a
chelating reagent such as EDTA Enzyme activities
were determined with or without EDTA
First, the caseinolytic activity of ovarian fluid was
examined The ovarian fluid was prepared from female
fish carrying embryos at the following stages: stages of
late blastula (stage 11), 22–23 somites (optic cups,
stage 20), auditory placodes (stage 21), 26–27 somites
(pectoral fins, stage 24), pigmentation of retina
(stage 25), openings of mouth and anus (stage 28),
pig-mentation of peritoneal wall (stage 29), depletion of
yolk (stage 30), immediately before hatching (stage 31),
and after embryo delivery [16] As shown in Fig 1A, constant activities were observed in the ovarian fluids carrying stage 11 to stage 30 embryos (stage 11 to stage 30 ovarian fluid) The activity was sharply increased in the stage 31 ovarian fluid, and disap-peared from the fluid after embryo delivery The activi-ties in stage 11 to stage 30 ovarian fluid were not inhibited by EDTA, but the activity in stage 31 ovar-ian fluid dropped to about a half because of EDTA Although some proteases are present in ovarian fluid carrying embryos throughout all developmental stages, the stage 31 ovarian fluid is suggested to contain metalloprotease(s)
Next, the substrate specificity of the enzyme activity was examined using Suc-Leu-Leu-Val-Tyr-7-amino-4-methylcoumarin (MCA) and Suc-Ala-Pro-Ala-MCA as substrates; these are the best substrates for medaka HCE [12] and Fundulus HCE [8], respectively Fig-ure 1B shows the change in MCA-peptide-cleaving activity of the ovarian fluid towards Suc-Leu-Leu-Val-Tyr-MCA Little or no activity was observed in stage 11 to stage 30 ovarian fluid The activity was sharply increased in the stage 31 fluid, and was not detected in the ovarian fluid after embryo delivery The activity in the stage 31 fluid was strongly inhibited
by EDTA The activity towards Suc-Ala-Pro-Ala-MCA in stage 31 ovarian fluid was about 30 times less than that towards Suc-Leu-Leu-Val-Tyr-MCA The changes in the activities throughout development were the same as those towards Suc-Leu-Leu-Val-Tyr-MCA These results suggest that the
metallo-Fig 1 Caseinolytic activity (A) and Suc-Leu-Leu-Val-Tyr-MCA-cleav-ing activity (B) of ovarian fluid carrySuc-Leu-Leu-Val-Tyr-MCA-cleav-ing embryos at various develop-mental stages (from stage 11 to stage 31) and after embryo delivery, D Black circles and white squares indicate the activities
of the fluid preincubated without and with 20 m M EDTA, respec-tively Caseinolytic and MCA-cleaving activities are expressed as
DA 280 30 min)1and nmolÆmin)1, respectively.
Trang 3protease with the substrate specificity similar to that of
known HCEs is present specifically in the stage 31
ovarian fluid
Choriolytic activity in stage 31 ovarian fluid and
morphological changes of the chorion
As stage 31 of black rockfish embryos is the stage
immediately before hatching, it is conceivable that
metalloprotease(s) present in the stage 31 ovarian fluid
are the hatching enzyme(s) of black rockfish When the
stage 31 ovarian fluid was incubated with chorion
frag-ments, the amount of liberated peptides was increased
up to 30 min and became constant thereafter
(Fig 2A) Most of the peptides were not liberated after
the treatment with EDTA, suggesting that metallopro-tease efficiently digesting the chorion is present in the stage 31 ovarian fluid After 30 min of incubation, the chorion was swollen (Fig 2D), and the thickness of the chorion was increased about four times when com-pared with that of the control chorion (Fig 2B,C) Eighty minutes later, the inner layer of the chorion was completely digested, and the thin outer layer remained undigested (Fig 2E)
The fine structure of the black rockfish chorion before
or after incubation with ovarian fluid was observed with
an electron microscope The control chorion was com-posed of a thick inner layer and a thin outer layer The inner layer seems to be composed of two layers, which are morphologically distinct (Fig 3A) No significant change of the chorion was observed after the incubation with stage 24 ovarian fluid (data not shown) On the other hand, stage 31 ovarian fluid swelled both of the inner layers of the isolated chorion (Fig 3B), and fine fibrillar structures were observed in the outer region of the inner layer (Fig 3C) This structural change was similar to that of the chorion isolated from stage 31 embryos (Fig 3D) The chorion-digesting property of the stage 31 ovarian fluid was similar to that of HCEs that have been previously reported in medaka and Fund-ulus [8,13] This observation suggests that an HCE-like activity, rather than an LCE-like activity, exists in stage 31 ovarian fluid
Identification of HCE from stage 31 ovarian fluid The protease(s) in stage 31 ovarian fluid was par-tially purified by successive HPLC steps through a gel
Fig 2 (A) Time course of chorion solubilization by stage 31 ovarian
fluid Black circles and white squares indicate the activities of the
fluid preincubated without and with 20 m M EDTA, respectively The
activity is expressed as the value of DA 595 Black rockfish chorion
isolated from stage 11 embryos was incubated for 0 min (B, C),
30 min (D) and 80 min (E) Scale bars: 100 lm Arrows indicate
thickness of chorion.
Fig 3 Electron microscopic observation of morphological change
of the chorion by stage 31 ovarian fluid The chorion isolated from stage 11 embryos was incubated with only the buffer (A) and with stage 31 ovarian fluid (B) (C) High magnification of the part shown
in the box in (B) The bar indicates the outer layer (D) The chorion isolated from a stage 31 embryo Scale bars: 1 lm (A, B, D) and 0.5 lm (C).
Trang 4filtration column, S-Sepharose column and Source 15S
column Figure 4 shows the chromatogram of the
Source 15S column Most of the proteins were
adsorbed to the column, and the proteolytic activity
was eluted as two peaks just after a large protein peak
Then, the fraction containing the two peaks was
sub-jected to reversed-phase column chromatography The
five protein peaks thus obtained were analyzed by
SDS⁄ PAGE The major peak, containing a 23 kDa
protein, the molecular mass of which was anticipated
to be the molecular mass of other euteleostean HCEs,
was subjected to MALDI-TOF MS analysis (Fig 4)
The values (m⁄ z 22 789.68 and 23 075.27) were almost
identical to the relative molecular masses calculated
from two black rockfish HCE cDNAs (SsHCE1,
Mr= 22 584; SsHCE2, Mr= 23 056) cloned in the
present study (described later) These results strongly
suggest that the chorion-swelling activity in the
stage 31 ovarian fluid is responsible for the action of
HCEs, the genes of which are orthologous to those of
other euteleostean HCEs
Cloning of black rockfish hatching enzyme
cDNAs
It has been suggested that both HCE and LCE genes
are present in euteleostean fishes [10] However, only
HCE was identified in stage 31 ovarian fluid Whether
black rockfish possess both the HCE and LCE genes
or not remains unclear First, we performed cloning of
hatching enzyme cDNAs by RT-PCR and RACE PCR
from the RNA of black rockfish embryos As a result,
the 1009 bp and 1088 bp cDNAs were cloned from black rockfish embryos Figure 5 shows the phyloge-netic tree constructed from the previously cloned hatching enzyme cDNAs of fishes belonging to the Elopomorpha (Japanese eel) and the Euteleostei (medaka, Fundulus, fugu, and Tetraodon), together with the cDNAs cloned in the present study The tree clearly shows that euteleostean hatching enzymes are divided into HCE and LCE clades with high probabil-ity (92% for the maximum likelihood tree, 100% for the neighbor-joining tree, and 100% for the Bayesian tree) On the basis of the tree, the two cloned cDNAs were named black rockfish Seb schlegelii HCEs, SsHCE1 and SsHCE2
Fig 4 Elution pattern of cation exchange Source 15S
chromatogra-phy with a linear gradient from 0 to 1 M NaCl Solid line,
absor-bance at 280 nm; dashed line, Suc-Leu-Leu-Val-Tyr-MCA-cleaving
activity shown as nmolÆmin)1 The inset shows the MALDI-TOF MS
spectrum obtained from the major peak by RP-HPLC with the
range of m ⁄ z values from 21 716 to 24 768 Ions at m ⁄ z 22 789.68
and 23 075.27 were identified as the black rockfish HCE.
Fig 5 A 55% majority rule consensus phylogenetic tree structed by the maximum likelihood method The tree was con-structed using nucleotide sequences at the mature enzyme portion
of hatching enzymes of arowana (AwHE, AB276000), bony tongue (BtHE, AB360712), Japanese eel (EHE, AB071423–9), Fundulus (FHCE, AB210813; and FLCE, AB210814), medaka (MHCE, M96170; and MLCE, M96169), Tetraodon (TnHCE, AB246043; and TnLCE, AB246044), fugu (FgHCE, AB246041; and FgLCE, AB246042), stickleback (GaHCE, AB353108–9; and GaLCE, AB353110), Set guentheri (SgHCE, AB353105–6; and SgLCE, AB353107), H hilgendorfi (HhHCE, AB353102–3; and HhLCE, AB353104), and black rockfish (SsHCE, AB353099–100; and wSsLCE, AB353101) Numbers at the nodes indicate bootstrap val-ues for the maximum likelihood tree and neighbor-joining tree, and Bayesian posterior probabilities, shown as percentages.
Trang 5To obtain evolutionary information, we amplified
HCE genes from genomic DNAs of Helicolenus
hil-gendorfi and Setarches guentheri, which belong to the
same subfamily (Sebastinae) as that of black rockfish
[15] From both the species, SsHCE1 and SsHCE2
or-thologs (HhHCE1 and HhHCE2 for H hilgendorfi,
and SgHCE1 and SgHCE2 for Set guentheri) were
cloned (Fig 5) HCE (GaHCE1 and GaHCE2)
cDNAs were also cloned from the stickleback
Gaster-osteus aculeatus, belonging to the Gasterosteiformes
[15], which is an order different from the
Scorpaenifor-mes Both the orders belong to the same series, the Percomorpha
The amino acid sequences of HCEs deduced from the newly cloned cDNAs are shown in Fig 6A All
of them possessed two active site consensus sequences of the astacin family proteases: HExxHxx-GFxHExxRxDR (zinc-binding site) and SxMHY (methionine turn) [17–19] In addition, six cysteines, which are present in all of the previously cloned fish hatching enzymes [9], were conserved among them
Fig 6 (A) A multiple alignment of amino acid sequences of hatching enzymes White and black triangles indicate putative signal sequence cleavage sites and N-terminals of mature enzymes, respectively Arrows indicate intron insertion sites of LCE genes Identical residues are boxed Dashes represent gaps Two active site consensus sequences of the astacin family protease are given in dark (zinc-binding site) and light (methionine turn) gray boxes, and conserved cysteine residues are in black boxes (B) Exon–intron structures of black rockfish (wSsLCE), H hilgendorfi (HhLCE), Set guentheri (SgLCE) and stickleback (GaLCE) LCE and HCE genes The exons and introns are indicated
by boxes and solid lines, respectively Numbers in parentheses indicate intron phases.
Trang 6The gene structures of all the HCE genes were
deter-mined to be intron-less (Fig 6B), which is
characteris-tic of HCE genes [10] Southern blot analysis showed
that the SsHCE1 probe hybridized with at least four
EcoRI fragments of 4.4, 3.8, 3.4 and 3.2 kbp of black
rockfish genomic DNA (Fig 7A), indicating that the
black rockfish HCE gene is a multicopy gene, like
other euteleostean HCE genes examined so far [10]
As no LCE cDNA fragments were obtained from
the black rockfish by the above strategy, we employed
another strategy: that is, primers were generated from
the sequence of stickleback LCE (GaLCE) cDNA Six
different-size cDNAs (600–2 kbp) were cloned from
black rockfish embryos, and five of the six were the
transcripts that would be formed by abnormal splicing
(see later) The other one (929 bp, SsLCE1) was well
aligned with other known LCE cDNAs, but its ORF
was incomplete Thus, the black rockfish LCE gene is
transcribed, but the gene is not translated into a
func-tional protein The LCE gene is predicted to be a
pseudogene We named it black rockfish pseudo-LCE
gene (wSsLCE) These results support the finding from
the protein level experiment that only HCE activity,
not the cooperative activity of HCE and LCE, is
pres-ent in stage 31 ovarian fluid
LCE genes were cloned from H hilgendorfi
(HhLCE) and Set guentheri (SgLCE) Their ORFs
were predicted to be complete Figure 8 shows
nucle-otide and deduced amino acid sequences of
wSsLCE1 and HhLCE cDNAs The identity of the
nucleotide sequences of the ORF between them was
wSsLCE1 cDNA possessed a pretermination stop
codon due to nucleotide substitution of 262G to
262T, and a frameshift mutation due to one nucleo-tide deletion (288delA) (Fig 8)
The gene structure of wSsLCE was determined using the nucleotide sequence of wSsLCE1 cDNA The wSsLCE gene was composed of eight exons and seven introns; its structure, including the positions of exon– intron boundaries and intron phases, was the same as that of other euteleostean LCE genes (Fig 6B) [10] Southern blot analysis was performed using genomic DNA digested with BamHI, HindIII, ScaI and BglII The wSsLCE1 DNA probe hybridized with a single fragment in each digest (Fig 7B), suggesting that the wSsLCE gene is a single-copy gene, like other euteleos-tean LCE genes examined so far [10]
As described above, in addition to wSsLCE1 cDNA, five different-size cDNAs were cloned from black rock-fish embryos using primers designed from the 5¢-UTR and 3¢-UTR for wSsLCE1 cDNA The wSsLCE2 (724 bp) and wSsLCE3 (606 bp) cDNAs were shorter than wSsLCE1 cDNA (870 bp), whereas wSsLCE4 (1033 bp), wSsLCE5 (2036 bp) and wSsLCE6 (1852 bp) cDNAs were longer than wSsLCE1 cDNA (Fig 9A) wSsLCE2 and wSsLCE3 cDNAs lacked the entire region of exon 4 (146 bp) and exon 4⁄ 5 (264 bp)
of the wSsLCE gene, respectively Considering that the wSsLCE gene is a single-copy gene, wSsLCE2 and wSsLCE3 cDNAs are predicted to be the products resulting from exon skipping by aberrant splicing As the pretermination stop codon and the nucleotide dele-tion are present in exon 4, wSsLCE2 and wSsLCE3 cDNAs have complete ORFs However, their trans-lated products lack the N-terminal region of the mature enzyme encoded by exon 4, and are considered
to be nonfunctional On the other hand, wSsLCE4 and wSsLCE5 cDNAs possessed the entire intron 1 (163 bp) and intron 5 (1166 bp) sequences, respec-tively, showing cancellation of splicing of intron 1 and intron 5, respectively wSsLCE6 cDNA was 184 bp shorter than wSsLCE5 cDNA, due to partial deletion
of exon 5 and partial inclusion of intron 5 wSsLCE6 cDNA is considered to be the transcript that appears
as a result of imprecise splicing
As shown in Fig 9B, intron regions including the 5¢-splicing boundary of intron 5 also showed the simi-larity among the black rockfish, H hilgendorfi and Set guentheri When we focused on the 5¢-splicing con-sensus sequence (gtragt) [20], we found a G to A sub-stitution in the +5 site of the wSsLCE gene (gtragt to gtgaat), whereas those of the HhLCE and SgLCE genes were well conserved An experiment has demon-strated that +5 site mutation causes the exon skipping [21] These results suggest that the mutation found in the wSsLCE gene probably results in intron 5 being
Fig 7 Southern blot analysis of SsHCE1 (A) and wSsLCE (B)
genes The restriction enzymes are shown at the top Numbers on
the left refer to the positions of size markers.
Trang 7retained by the cancellation of splicing, as seen in
wSsLCE5 cDNA, and in the exon deletion, as seen in
wSsLCE3 cDNA (Fig 9A)
Half of the wSsLCE cDNAs cloned in the present
study had one nucleotide deletion (73delG) located at
the 5¢-end of exon 2 (Fig 8) The region including the
exon–intron boundary between intron 1 and exon 2
was amplified by PCR from the genomic DNA
Sequence analysis revealed that the gene is
heterozy-gous, and that a nucleotide substitution-destroying
splicing acceptor consensus sequence (AG to AA;
Fig 9B) is present in one of the alleleic wSsLCE genes
One of the alleles used the original AG acceptor
sequence, and the other mutated allele used a
pseudo-AG acceptor sequence by shifting one nucleotide to
the 3¢-site; that is, )1A in the intronic sequence and
73G in the exonic sequence were used as the acceptor
sites The occurrence of 73delG in wSsLCE cDNA can
be explained if the 73G was spliced out for use as a
pseudo-AG acceptor sequence (Fig 9B) The substi-tution might also cause the intron 1 retention, as seen
in wSsLCE4 cDNA (Fig 9A)
Expression of black rockfish hatching enzyme genes
First, the gene expression of SsHCE and wSsLCE was analyzed by northern blot analysis An SsHCE1 DNA probe was used for detecting the HCE transcript This probe probably detects both the SsHCE1 and SsHCE2 transcripts, because of their high level of similarity (88%) The hybridization of this probe with 10 lg of total RNA did not show any signal This amount of RNA, 10 lg, is known to be enough for detecting the HCE transcripts of medaka and Fundulus [8,22] The result suggests that the expression of SsHCE genes is much weaker than that in other fish species, and there-fore, poly(A)-rich RNA purified from 100 lg of total
Fig 8 Nucleotide and predicted amino acid sequences of wSsLCE1 and HhLCE Arrows indicate intron insertion sites with intron numbers Boxes indicate mutation sites found in the wSsLCE gene as described in the text.
Trang 8RNA was employed The SsHCE1 probe hybridized
with about 1 kb of transcript; this size was consistent
with that of the cDNAs The transcripts were detected
in stage 17⁄ 18 embryos, decreased in amount towards
stage 25, and disappeared thereafter (Fig 10A) We
failed to detect the positive signal of the wSsLCE gene
transcript by northern blot analysis
Next, gene expression was determined by RT-PCR
(Fig 10B) After 28 cycles of PCR, sufficient
expres-sion of the SsHCE1 and SsHCE2 genes was
detected, and the band intensity of SsHCE2
tran-scripts was about half that of SsHCE1 For the
wSsLCE gene, the 33 cycles of RT-PCR gave these
bands at about 700 bp, 800 bp, 1 kbp, and 1.2 kbp,
corresponding to wSsLCE3, wSsLCE2, wSsLCE1 and
wSsLCE4 cDNAs, respectively The expression
pat-tern of the wSsLCE gene through the developmental
stages was similar to that of the SsHCE genes, but
the expression was much weaker than that of the SsHCE genes
As shown in Fig 11, whole-mount in situ hybrid-ization using an antisense RNA probe for the SsHCE1 gene revealed a distribution of cells express-ing SsHCE transcripts in developexpress-ing black rockfish embryos It is well known that the fish hatching gland cells differentiate at the anterior end of the hypoblast layer, called the pillow, in the late gastrula embryos, and until hatching, the gland cells migrate
to the final destination in a species-dependent man-ner [5,22] In stage 17 embryos of the black rockfish, positive cells were first observed along the edge of the anterior head These cells seem to make a start
in migration from the pillow (Fig 11A) From stage 18 to stage 22, the cells migrated posteriorly (Fig 11B), and they were finally distributed widely
in the epidermis of both lateral sides of the head
Fig 9 (A) A schematic representation of the splicing variants of the wSsLCE gene The black triangle indicates putative N-terminals of mature enzymes The structures of the normally spliced form (wSsLCE1) and the alternatively spliced forms (wSsLCE2–6) are shown wSsLCE2, wSsLCE3, wSsLCE4, wSsLCE5 and wSsLCE6 have an exon 4 deletion, an exon 4 and 5 deletion, an intron 1 inclusion, an intron 5 inclusion, and partial deletion of exon 5 and partial inclusion of intron 5, respectively (B) Nucleotide mutations found on the splice site con-sensus sequence at intron 5 and intron 1 The upper part gives a comparison of the exon–intron boundary between exon 5 and intron 5 among the wSsLCE, HhLCE and SgLCE genes The consensus sequence of splicing donor site is shown at the top The lower part is an electropherogram of the PCR product around the boundary between intron 1 and exon 2 The splicing acceptor consensus sequence and pseudo-AG consensus sequence are indicated by red boxes on the upper and lower lines, respectively, together with each cDNA product The regions of the exon and intron are indicated by upper-case and lower-case letters, respectively.
Trang 9(Fig 11C,D) In stage 24 and stage 25 embryos, the
signals in positive cells became weak and their
num-bers were decreased No signals were observed in
stage 29 and stage 31 embryos and posthatching fry, and nor were signals from sense RNA observed in any embryos
Fig 10 Expression analysis of the SsHCE1, SsHCE2 and wSsLCE genes (A) Northern blot analysis of expression of the SsHCE gene during development Arrowheads indi-cate the positions of 28S and 18S rRNA (B) RT-PCR analysis of SsHCE1, SsHCE2 and wSsLCE during development b-Actin was used as a control PCR cycles were 28 for SsHCE1 and SsHCE2, 33 for wSsLCE, and
24 for b-actin Developmental stages are shown at the top Fry, posthatching embryos The 200 bp (SsHCE1, SsHCE2, and wSsLCE) and 100 bp (b-actin) ladder markers are shown in the left lane.
Fig 11 Whole-mount in situ hybridization
of SsHCE gene during the development of black rockfish embryos The SsHCE1 RNA probe was hybridized with stage 17 (A), stage 18 (B), stage 22 (C, D), stage 24 (E) and stage 25 (F) embryos (A, B) Dorsal views of head regions Upper, the anterior-most (C, E, F) Lateral views Upper, dorsal (D) Dorsal view of the head region Right, the anterior-most Yolk was removed from stage 22 embryos (C, D) Scale bars:
200 lm (G) Average number of hatching gland cells per embryo The values are expressed as the mean of five embryos Error bars indicate the standard deviation.
Trang 10Throughout the developmental stages, the total
number of SsHCE-expressing cells per embryo seemed
to be less than in other fishes The number of hatching
gland cells in hybridized embryos was counted, and
the average number per embryo was determined at
each developmental stage (Fig 11G) In stage 17 and
stage 22 embryos, about 100 cells were observed, and
the number was decreased to about one-half at
stage 24, to about one-quarter at stage 25, and to zero
at stage 29 These results were consistent with the
developmental expression profile obtained by northern
blot analysis In comparison, we counted the numbers
of hatching gland cells of rainbow trout, ayu or loach
embryos at the middle to late stages of somitogenesis
There were about 3000 (loach), 2000 (rainbow trout)
and 1000 (ayu) per embryo Thus, black rockfish
hatching gland cells were about 10–30 times fewer in
number than those of other fish species Summing up
the results, the black rockfish hatching enzyme gene is
actively expressed, but its expression stops at the
ear-lier stages In addition, the expression level is
consid-ered to be suppressed to a greater extent than in other
fishes
Discussion
We investigated the hatching of an ovoviviparous
black rockfish The EDTA-sensitive protease activity
with a substrate specificity similar to that of known
HCEs was detected in the ovarian fluid carrying
embryos immediately before hatching stage (stage 31)
Furthermore, the protease was found to swell the inner
layer of the egg envelope (chorion) and to release some
water-soluble peptides from the chorion HCE, one of
the euteleostean hatching enzymes, is well known to
swell the chorion by its proteolytic action The
prote-ases in the stage 31 ovarian fluid were partially
puri-fied, and a proteolytically active fraction containing
proteins had a molecular mass corresponding to the
cloned SsHCE1 and SsHCE2 cDNAs according to
MALDI-TOF MS analysis Therefore, these results
strongly suggest that HCEs are secreted from black
rockfish embryos immediately before the hatching
stage This is the first demonstration of hatching
enzymes in ovoviviparous fish
At the natural hatching of medaka and Fundulus
embryos, the chorion is efficiently solubilized, and no
swelling of the chorion has been observed, due to the
concurrent and cooperative action of LCE and HCE
[8,13] The morphological change of the chorion
observed in black rockfish embryos implies that its
chorion digestion mechanism is different from that of
other euteleostean fishes In addition, the present study
revealed that HCE cDNAs were cloned and their gene expression was observed specifically in the hatching gland cells of embryos, whereas the LCE gene was pseudogenized These results suggest that the chorion digestion at black rockfish hatching is performed by HCE alone The intact chorion of the black rockfish was thin and fragile when compared with the medaka and Fundulus chorions (Fig 2B), and had about one-fourth the thickness of the medaka chorion [23] According to in vitro experiments, the chorion was completely digested by a long period of incubation (80 min) with stage 31 ovarian fluid Considering that the hatching enzyme stays with the chorion for a long time in the ovarian cavity, HCE alone would be suffi-cient for chorion digestion
The northern blot analysis and in situ hybridization experiment showed that expression of the HCE gene was suppressed to a very low extent when compared with that of other euteleostean HCE genes In addi-tion, the hatching enzyme synthesis of the black rock-fish ceased around the middle of somitogenesis, whereas that of other teleostean fishes, such as medaka, zebrafish, Japanese eel and ayu, could be detected at stages from the beginning of its expression
to immediately before hatching [5,7,9,22] These results imply that the black rockfish embryo synthesizes an amount sufficient for, but limited to, chorion digestion Such an amount would not be harmful for embryos,
as embryos might be damaged by a long period of incubation with a high concentration of the protease Thus, the hatching enzyme system in oviparous fish embryos is conserved in the ovoviviparous black rockfish, with adaptations to their specific hatching environment
According to the teleostean phylogenetic tree pro-posed by Nelson, the ovoviviparous black rockfish and oviparous H hilgendorfi belong to the same tribe (Sebastinae) but different genera, and oviparous Set guentheri belongs to the same subfamily (Sebasti-nae) but a different tribe [15] The mitochondrial DNA-based phylogenetic tree indicates that the genus Helicolenus is sister to Sebastes, which includes the black rockfish [24] The nucleotide sequences of black rockfish hatching enzyme cDNAs indicated high simi-larity (93% and 97% for HCE1 and HCE2, respec-tively, and 95% for LCE) to those of H hilgendorfi, and the phylogenetic analysis (Fig 5) agreed well with the mitochondrial phylogenetic tree Despite this phy-logenetically close relationship, the LCE genes of
H hilgendorfi and Set guentheri had complete ORFs, whereas that of the black rockfish was incomplete The Sebastesfossils can be traced back to the late Miocene (about 6–10 million years ago, MYA) [25] This time